Machine Learning
Mohammad Reza Nazabadi; Seyed Esmaeil Najafi; Ali Mohaghar; Farzad Movahedi Sobhani
Abstract
Adopting an integrated production, maintenance, and quality policy in production systems is of great importance due to their interconnected influence. Consequently, investigating these aspects in isolation may yield an infeasible solution. This paper aims to address the joint optimal policy of production, ...
Read More
Adopting an integrated production, maintenance, and quality policy in production systems is of great importance due to their interconnected influence. Consequently, investigating these aspects in isolation may yield an infeasible solution. This paper aims to address the joint optimal policy of production, maintenance, and quality in a two-machine-single-product production system with an intermediate buffer and final product storage. The production machines have degradation levels from as-good-as-new to the breakdown state. The failures increase the production machine's degradation level, and maintenance activities change the status to the initial state. Also, the quality of the final product depends on the level of degradation of the machines and the correlation between the degradation level of the production machines and the product's quality in the case that high degradation of the previous production machines leads to a high probability to produce wastage by the following machines is considered. The production system studied in this research has been modeled using the agent-based simulation, and the Reinforcement Learning (RL) algorithm has obtained the optimal integrated policy. The goal is to find an integrated optimal policy that minimizes production costs, maintenance costs, inventory costs, lost orders, breakdown of production machines, and low-quality production. The meta-heuristic technique evaluates the joint policy obtained by the decision-maker agent. The results show that the acquired joint policy by the RL algorithm offers acceptable performance and can be applied to the autonomous real-time decision-making process in manufacturing systems.
Machine Learning
Parvaneh Afzali; Abdoreza Rezapour; Ahmad Rezaee Jordehi
Abstract
The authentication of writers through handwritten text stands as a biometric technique with considerable practical importance in the field of document forensics and literary history. The verification process involves a meticulous examination of the questioned handwriting in comparison to the genuine ...
Read More
The authentication of writers through handwritten text stands as a biometric technique with considerable practical importance in the field of document forensics and literary history. The verification process involves a meticulous examination of the questioned handwriting in comparison to the genuine handwriting of a known writer, aiming to determine whether a shared authorship exists. In real-world scenarios, writer verification based on the handwritten text presents more challenges compared to signatures. Signatures typically consist of fixed designs chosen by signers, whereas textual content can vary and encompass a diverse set of letters, numbers, and punctuation marks. Moreover, verifying a writer based on limited handwritten texts, such as a single word, is recognized as one of authentication's open and challenging aspects. In this paper, we propose a Customized Siamese Convolutional Neural Network (CSCNN) for offline writer verification based on handwritten words. Additionally, a combined loss function is employed to achieve more accurate discrimination between the handwriting styles of different writers. The designed model is trained with pairs of images, each comprising one authentic and one questioned handwritten word. The effectiveness of the proposed model is substantiated through experimental results obtained from two well-known datasets in both English and Arabic, IAM and IFN/ENIT. These results underscore the efficiency and performance of our model across diverse linguistic contexts.
Machine Learning
H. Herunde; A. Singh; H. Deshpande; P. Shetty
Abstract
Nowadays, the control of the traffic in the urban roads and in the highway has been a big challenge as the number of increase in the auto mobiles. So to overcome this problem we use the detection and tracking the vehicles using the traffic surveillance system. We can manage and control the traffic more ...
Read More
Nowadays, the control of the traffic in the urban roads and in the highway has been a big challenge as the number of increase in the auto mobiles. So to overcome this problem we use the detection and tracking the vehicles using the traffic surveillance system. We can manage and control the traffic more easily. It is very complicated and a challenging task to identify the vehicle or a moving object in a complex environment with various background. The ratio detected of such algorithms depends on the quality of the foreground mask generated. Therefore this project is to present the detection and tracking the vehicles and the pedestrians in an efficient method which focus on trajectory motion of the vehicles and the pedestrians. In this proposed method, the pixels in the background are preserved which can be cars, bikes, buses, pedestrian, etc., the rest is discarded as the noise. Hence, our proposed method detects the vehicles and the pedestrians as mentioned and discards the rest noise as well in the same time. Here the quality of the generated foreground mask is more to increase the detection ratio. The performance is compared with other standard methods qualitatively and quantitatively.
Machine Learning
A. Singh; H. Herunde; F. Furtado
Abstract
Amid the previous three decades, the topic of image processing has gained vital name and recognition among researchers because of their frequent look in varied and widespread applications within the field of various branches of science and engineering. As an example, image processing is helpful to issues ...
Read More
Amid the previous three decades, the topic of image processing has gained vital name and recognition among researchers because of their frequent look in varied and widespread applications within the field of various branches of science and engineering. As an example, image processing is helpful to issues in signature recognition, digital video processing, remote sensing and finance. Image processing models are used for detecting the face. The aim of this thesis is to solve the face-detection in the first attempt using the Haar-cascade classifier from images containing simple and complex backgrounds. It is one of the preeminent detectors in terms of reliability and speed. We introduced a new method to deal with the frontal face images by using a modified Haar cascade algorithm. By using this algorithm, we can detect the image as well as the coordinates. The main attraction of this paper is to solve different types of images having one object, two objects, and three objects which can’t be solved by any of the existing methods but can be solved by our proposed method.
Machine Learning
H. Deshpande; A. Singh; H. Herunde
Abstract
Computer Vision is a field of study that helps to develop techniques to identify images and displays. It has various features like image recognition, object detection and image creation, etc. Object detection is used for face detection, vehicle detection, web images, and safety systems. Its algorithms ...
Read More
Computer Vision is a field of study that helps to develop techniques to identify images and displays. It has various features like image recognition, object detection and image creation, etc. Object detection is used for face detection, vehicle detection, web images, and safety systems. Its algorithms are Region-based Convolutional Neural Networks (RCNN), Faster-RCNN and You Only Look Once Method (YOLO) that have shown state-of-the-art performance. Of these, YOLO is better in speed compared to accuracy. It has efficient object detection without compromising on performance.
Machine Learning
F. Furtado; A. Singh
Abstract
Nowadays, the recommendation system has made finding the things easy that we need. Movie recommendation systems aim at helping movie enthusiasts by suggesting what movie to watch without having to go through the long process of choosing from a large set of movies which go up to thousands and millions ...
Read More
Nowadays, the recommendation system has made finding the things easy that we need. Movie recommendation systems aim at helping movie enthusiasts by suggesting what movie to watch without having to go through the long process of choosing from a large set of movies which go up to thousands and millions that is time consuming and confusing. In this article, our aim is to reduce the human effort by suggesting movies based on the user’s interests. To handle such problems, we introduced a model combining both content-based and collaborative approach. It will give progressively explicit outcomes compared to different systems that are based on content-based approach. Content-based recommendation systems are constrained to people, these systems don’t prescribe things out of the box, thus limiting your choice to explore more. Hence, we have focused on a system that resolves these issues.
Machine Learning
A. Ghanbari Talouki; M. Majdi
Abstract
Inpainting or completion is used with the purpose of restoring damaged images and video frames. This paper proposes an applicable algorithm to inpaint corrupted subjects in video frames. To begin with, background and foreground (moving subject) are separated from each other in each frame, with the aim ...
Read More
Inpainting or completion is used with the purpose of restoring damaged images and video frames. This paper proposes an applicable algorithm to inpaint corrupted subjects in video frames. To begin with, background and foreground (moving subject) are separated from each other in each frame, with the aim of getting to a more visually pleasant result. Static background inpainting is done using a patch-based method. To inpaint the corrupted moving subject, a subject-based method that is an improvement on the rigid object-based method is used in this study to consider the special issue of inpainting the human body. To fill in the holes created by occluding subjects the most appropriate template is found using a similarity measure which is based on both contour and pixel values. The inpainted video is acquired by superimposing the completed foreground on the inpainted background